INDEX

Explanations

instances where phrases mention groups of people or time periods

oai_token-act-pair · gpt-3.5-turbo

phrases indicating common experiences or sentiments shared by many people

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bil

-0.72

obin

-0.69

Tro

-0.63

mosp

-0.63

 foremost

-0.62

 steal

-0.60

 rout

-0.60

 sequ

-0.58

gart

-0.58

bid

-0.57

POSITIVE LOGITS

 sake

1.24

 purposes

1.17

 reasons

0.89

ummies

0.89

 occasions

0.79

 foreseeable

0.77

 eternity

0.72

 duration

0.71

ulz

0.70

 households

0.68

Activations Density 0.080%

instances where phrases mention groups of people or time periods

phrases indicating common experiences or sentiments shared by many people

No Comments

No Known Activations